Darwin-Katherine

The demand forecast is used to predict how the power demand will change in the near future. An explanation of the different forecasts is given below:

• Demand is the actual demand at the indicated time.
• Forecasts are the predictions of what the demand is likely to be at the indicated time. These predictions are made using a machine learning model with the current demand as well as current and predicted future weather conditions as inputs. Three types of forecasts are given:
  • Demand Forecast is the best guess for what the demand will be at the indicated time.
  • P10 is a lower bound for the expected demand at the indicated time. The name means that the demand has a 10% chance of being below this forecast, and a 90% chance of being above it.
  • P90 is an upper bound for the expected demand at the indicated time. The name means that the demand has a 90% chance of being below this forecast, and a 10% chance of being above it.
• Historic Forecasts give an indication of how well the forecast has been performing. The data point for any indicated time shows what it was forecasted to be 30 minutes before that time. For example, if the “Historic p90 Forecast” is 227 MW at 10:45 am, it means that at 10:15 am, the “p90 Forecast” would have shown 227 MW for 10:45 am.

The forecasts are developed, in part, from information provided to NTESMO by third parties and NTESMO does not verify the accuracy of this information. As forecasts of this nature are inherently subject to error and degrees of uncertainty, no assumption can be made that actual events will reflect a forecast.